search

clustervision / trinityX

TrinityX is the new generation of ClusterVision's open-source HPC, A/I and cloudbursting platform. It is designed from the ground up to provide all services required in a modern HPC and A/I system, and to allow full customization of the installation.
GNU General Public License v3.0
67 stars 37 forks source link

two issue regarding controller.yml - OpenLDAP start service #429

Open jameslam0250 opened 3 months ago

jameslam0250 commented 3 months ago

first , if the installer re-runs , the consine etc ldif may have doubled entry and need to think if the openldap-servers are required to uninstall prior to install again.

I was trying to install using the INSTALL.sh and we meet the systemctl start slapd timeout during the ansible playbook. issue investigating.

aphmschonewille commented 3 months ago

Hi James,

would you have the output of the spot where it went wrong, say the last 50 lines of the playbook output? Info regarding your OS, release and rpm package list (rpm -qa --last) would be helpful. Though we ourselves re-run whole playbooks (while testing other parts), we did not see this one before.

-A

jameslam0250 commented 3 months ago

We will try again using my environment first. As the testing environment has proxy. Rocky 9.4 from yesterday testing site

On Wed, Aug 21, 2024, 4:21 PM aphmschonewille @.***> wrote:

Hi James,

would you have the output of the spot where it went wrong, say the last 50 lines of the playbook output? Info regarding your OS, release and rpm package list (rpm -qa --last) would be helpful. Though we ourselves re-run whole playbooks (while testing other parts), we did not see this one before.

-A

— Reply to this email directly, view it on GitHub https://github.com/clustervision/trinityX/issues/429#issuecomment-2301449932, or unsubscribe https://github.com/notifications/unsubscribe-auth/AJMAQLGIM5NBH5IVOFXZUM3ZSREZHAVCNFSM6AAAAABM2BY2TWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMBRGQ2DSOJTGI . You are receiving this because you authored the thread.Message ID: @.***>

jameslam0250 commented 3 months ago

OK , I just tested with my faster site and openldap issue is gone. but one thing that I need to ask is :

if their network is closed (or even limited bandwidth) and they are heavily using local ISO do deploy the nodes. What could I do to fix the repository of the rocky Appstream / BaseOS to local directory -that contain all content of repo-

As I can see the bandwidth usage during the ansible is quite high. ( my pre-production site do only have 2MB/s at most)

aphmschonewille commented 3 months ago

For some time now we are planning to make an airgapped install method (ISO), but this is not an easy feat as we use many other resources like github repos and pip. When (not if) that's available, we announce it. I expect not sooner than late october.

jameslam0250 commented 2 months ago

OK, we will wait you for that , but for compute node image trinity.log

we would like to ask if we can skip the online requesting mirror.rockylinux.org and point to file:///distro as my baseurl of the compute node image. (as I mentioned , bandwidth restricted)

aphmschonewille commented 1 month ago

the airgapped release will take longer than expected due to the complexity of containing external (git)repos, downloads and inclusions. Maybe there is an alternative for you in the meantime? We support the use of a proxy, which helps us at least during the many tests and pipelines. Maybe this is something you could consider?

jameslam0250 commented 1 month ago

Actually we were using proxy as alternative but the Ansible stops frequently at compute node establish and even for master node setup

That was a pain when setup using proxy and we need to fix the IP address under /etc/host to fix the connection issue

However, this doesn't work for compute node image as the image building relies on mirror site

And that's why we would like to ask-

Where can we change the path of the mirror site to file:/// as alternative option. Since shooting a ISO inside is much easier than using the internet

On Wed, Oct 23, 2024, 3:45 PM aphmschonewille @.***> wrote:

the airgapped release will take longer than expected due to the complexity of containing external (git)repos, downloads and inclusions. Maybe there is an alternative for you in the meantime? We support the use of a proxy, which helps us at least during the many tests and pipelines. Maybe this is something you could consider?

— Reply to this email directly, view it on GitHub https://github.com/clustervision/trinityX/issues/429#issuecomment-2431177222, or unsubscribe https://github.com/notifications/unsubscribe-auth/AJMAQLCNCTDJW2KN4MWEUXDZ45HZTAVCNFSM6AAAAABM2BY2TWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIMZRGE3TOMRSGI . You are receiving this because you authored the thread.Message ID: @.***>